Naver's DAN25: Dual AI Strategy for Services and Industry
Naver set a clear course at its unified TeamNaver conference, DAN25: scale AI agents across core services, and push AI transformation in heavy industry. For executives, this is a signal to re-architect around agent-driven journeys, first-party data, and vertical AI.
Pillar 1: Agent N across Search, Shopping, Ads, and Business Operations
Naver will roll out an AI shopping agent in Q1 on the Naver Plus Store, followed by an "AI Tab" in Q2 that brings agentic, conversational search to the main experience. Users will chat with Agent N, which will interpret intent, retrieve the right content or product, and execute actions like purchase and payment.
Under the hood, Agent N sits on Naver's On-service AI, tying together service data and context to predict next steps and complete end-to-end tasks. Naver highlighted a data advantage: verified-buyer reviews, seller-linked inventory, and technical validation pipelines to keep agent actions reliable.
For businesses, "Agent N for Business" pulls fragmented tools across shopping, ads, and local into a single AI hub. The goal: diagnose performance, recommend fixes, and implement changes without manual swivel-chair work.
Pillar 2: AX for Manufacturing and "Physical AI"
Naver is targeting AI transformation in semiconductors, automobiles, and shipbuilding, led by Naver Cloud's vertical AI approach based on "Sovereign AI 2.0." The company is building industry-specific models and collaborating with leaders in shipbuilding, energy, and biotech, with expansion planned for Saudi Arabia, Thailand, and Japan.
"Physical AI" is being tested across real industrial environments, including defense and manufacturing. To support scale, Naver will invest over 1 trillion KRW in GPUs by 2026 and operate a testbed linking its 1784 HQ and the Sejong data center.
Why this matters for executives
- Search and shopping shift to conversations: Zero-query experiences will change funnel design, attribution, and merchandising. Prepare for agent-led paths, fewer clicks, and action-level metrics.
- First-party data becomes your edge: Verified reviews, inventory, and customer context feed better agent decisions. Close gaps in product data, availability, and policy rules now.
- Marketing automation gets smarter: Business agents will recommend and implement changes across ads and placements. Teams move from manual tuning to policy and guardrail design.
- Industrial AI goes vertical: Sector-specific models and "Physical AI" reduce error rates and downtime. Expect vendors to ask for tighter data access, process mapping, and safety cases.
- Sovereignty and privacy rise in priority: Data residency and model governance matter, especially for regulated sectors and cross-border work.
90-day action checklist
- Inventory your data: Product attributes, verified reviews, inventory feeds, service policies, and PII handling. Fix missing fields and stale syncs.
- Define 3-5 agent use cases: Example: guided product selection, returns automation, warranty claims, ad budget reallocation, or supplier ETA updates.
- Design guardrails: What agents can say, do, refund, or change. Add human-in-the-loop for higher-risk actions.
- Pilot with clear KPIs: Target conversion lift, AOV, issue resolution time, containment rate, and cost per resolved task.
- Prep for the AI Tab: Consolidate content and product data; structure FAQs and specs for agent retrieval; align landing pages with conversational intent.
- Marketing workflow: Test Naver Business Agent for pacing, bid, and creative suggestions. Track ROAS, time-to-launch, and error rate.
- Industrial track: Map a single process (inspection, scheduling, or quality). Instrument data capture, set safety constraints, and run a sandbox trial.
- Governance: Update data-sharing agreements, incident response, and model audit trails. Clarify who approves prompts, policies, and changes.
- Upskill your leads: Train product, ops, and marketing managers on agent design, prompts, and measurement. Practical AI training by job role.
Metrics to track
- Agent performance: Containment rate, action success rate, latency to decision, escalation rate, CSAT.
- Commerce: Conversion lift, AOV, return rate, out-of-stock avoidance, time-to-purchase.
- Ads: ROAS, CPA/CAC, creative iteration time, waste reduction from invalid traffic.
- Industrial: Defect rate delta, unplanned downtime, lead time, schedule adherence, incident count.
- Governance: Data residency compliance, access violations, model change logs, audit completion.
Risk controls and how Naver is addressing them
- Data quality: Verified-buyer reviews and seller-linked inventory reduce noise and misinformation.
- Execution safety: End-to-end validation pipelines help prevent wrong actions during purchase and fulfillment.
- Vertical models: Industry-specific AI narrows scope, improving accuracy for manufacturing use cases.
- Sovereign approach: Local language, data, and residency options support compliance-heavy sectors.
- Physical AI: Real-world testbeds across 1784 and Sejong aim to prove reliability before broad rollout.
Key dates and next steps
- Q1: AI shopping agent on Naver Plus Store.
- Q2: AI Tab for agentic, conversational search.
- Through 2026: >1 trillion KRW in GPUs; scaling industrial pilots and "Physical AI."
- Global expansion: Industry deployments targeted for Saudi Arabia, Thailand, and Japan.
- Creators and partners: AI/XR content programs and partnerships across gaming, music, and OTT; Naver Impact to support SMEs and local businesses.
What to do now
Pick one customer journey and one plant process. Wire in data, define guardrails, and run a four-week pilot with weekly KPI reviews. Small wins compound once the agent sees real traffic.
Learn more about the companies leading this push: Naver Corp and Naver Cloud.
Your membership also unlocks:
